Interactive Document Classification with Conflict Labeled Instances
Speaker: Ding-Yi Chen, ITEE
When: 2004-06-23 14:00:00
Venue: 78-420
Host: Dr. Xue Li
Abstract:We observed that in interactive text classification, user tends to
point out only the misclassified documents, not the correct ones. It
is unlikely that a user would be diligent enough to identify all the
misclassified documents. In this case, a classifier is expected to
deal with misclassified documents. Among them it is possible that
only a small proportion has been identified. We propose the
Prediction-Learning-Distillation (PLD) framework for distilling the
misclassified documents. Whenever a user points out an error, the
PLD learns from the mistake and identifies the same mistake from all
other classified documents. The PLD then enforces this learning for
future classifications. Our experiment results have demonstrated
that the proposed algorithm can learn from user identified
misclassified documents, then distills the rest successfully.
However, we also found that PLD can discover the misfiled documents,
but sometimes it will fail to assign labels correctly. To address
this problem, we discover that conflict labeled instances can be
used to explain this situation. Conflict labeled instances also have
several interesting usages, such as telling the upper bound of
effectiveness of classifier, fitness of document model, and they can
even be used to get a classifier. We use distribution hyperplane,
which is calculated from conflict labeled instances, in order to
build a classifier, know the theoretical upper bound of
effectiveness of classifier, and fitness of document model.
Biography:Ding-Yi Chen graduated from the National Central University (Taiwan)
with a degree of Bachelor of Information Management in 1994. He also
completed an Master degree in Information Management at National
Chi-Nan University (Taiwan) in 1999.
Type: Ph.D confirmation
Contact:Dr. Xue Li, seminar host (xueli@itee.uq.edu.au)
or Guido Governatori (ITEE seminar co-ordinator)
(guido@itee.uq.edu.au)